Re: CVS commit: pkgsrc/x11/qt4-libs

[Greg Troxel <gdt%lexort.com@localhost>]

Frédéric Fauberteau <triaxx%NetBSD.org@localhost> writes:

> Le 15/09/2023 à 13:14, Tobias Nygren a écrit :
>> On Wed, 13 Sep 2023 18:20:13 -0400
>> Greg Troxel <gdt%lexort.com@localhost> wrote:
>> 
>>> Frédéric Fauberteau <triaxx%netbsd.org@localhost> writes:
>>>
>>>> +NOT_FOR_PLATFORM= Darwin-*-aarch64
>>>
>>> Please switch to BROKEN_ON.  The general expectation is that software is
>>> portable, and there is no good reason why qt4 should be unavailable on
>>> this platorm.
>> The asm error is easy to fix with a patch, but qt4-libs won't work
>> anyway on macOS newer than 13, even on x86_64, because of deprecated &
>> removed Cocoa font functions. This is unlikely to ever be fixed and
>> meets the requirement of NOT_FOR_PLATFORM+=Darwin-*-* as the platform
>> did remove required things.
>> But it should perhaps be qualified with an additional
>> OPSYS_VERSION >= 130000 condition.
>> -Tobias
>
> Do you agree?
> diff -r1.56 Makefile.common
> 22c22,24
> < BROKEN_ON_PLATFORM=   Darwin-*-aarch64
> ---
>> .if ${OPSYS_VERSION} >= 130000
>> NOT_FOR_PLATFORM+=    Darwin-*-*
>> .endif

I don't agree with NOT_FOR.  I don't see any fundamental reason why qt4
could not be made to work, using some interfaces.  The fact that the
platform withdrew an interface that was used previously is not enough.
NetBSD doesn't have those interfaces, and qt4 works.  What's really
going on is that qt4 is not maintained and thus hasn't adapted, and
won't.  I bet qt5 and qt6 either work or will be made to work.

NOT_FOR_PLATFORM is supposed to be about things where it fundamentally
does not make sense, and there I would expect ONLY_FOR_PLATFORM much
more often, to limit to FooOS when the program is "utility to manipulate
the bar interface which is only present on FooOS".

But if it is only troubled on >= 13, then indeed limiting to that is good.